Tomographic reconstruction from non-calibrated noisy projections in non-destructive evaluation

We focus on three-dimensional reconstruction when a few two-dimensional noisy projections are available with limited angle projections. Furthermore, in this paper, the projections are not calibrated since they are given up to a multiplicative constant. With such incomplete data, classical techniques fail to produce acceptable reconstructions. It is well known that the introduction of regularization in the form of a cost function penalizing irregularity of the reconstructed object can improve performance. However, to obtain an accurate reconstruction, it is necessary to estimate the parameters of the cost function. In this paper, we address this problem.